Patient-Directed Computational Analysis of Atrial Fibrillation Project Summary Atrial fibrillation (AF) is the most common arrhythmia and is a rapidly-growing public health problem that currently affects over 30 million people world-wide and more than 5 million people in the US. If left untreated, AF leads to an increase in stroke, heart failure and mortality. Unfortunately, the mechanisms that maintain AF remain poorly understood, hindering further improvement of current therapy strategies. Recent studies, however, have revealed that AF may be driven by rotational or focal sources and that targeting these sources using localized ablation can result in promising long-term outcomes. Due to our incomplete understanding of AF, however, this targeted ablation approach is not always successful. This project will test the novel hypothesis that AF is sustained by localized rotational and focal sources with different size and temporal stability and that, after these sources are removed, termination is not immediate but is maintained by non-local mechanisms. We will address this hypothesis using a combined computational/clinical approach that employs advanced multiscale computational techniques and state-of-the-art clinical mapping. The project will 1) quantify AF organization using patient-specific geometries; 2) determine whether some rotational or focal sources are more important than others; 3) test possible causes of AF maintenance and termination using patient-specific digital computer models. We will use data from our unique and large patient registry, currently totaling >500 patients. This project is significant because it will establish a deeper understanding of AF and might reveal novel mechanisms of AF maintenance. Our results can be translated directly to practice and may enable the development of better treatment options.